Children begin producing language at about their first year. They slowly add new words to their vocabulary. Then towards the end of their second year, vocabulary growth accelerates --- averaging by some counts as many as nine new words a day. At this point, children often need to hear only one object labelled (one specific tractor called tractor for instance) to systematically generalize that name to other instances of the category. Past results show that young children generalize a novel object name to instances principally by attending to the object's shape. This research tests the hypothesis that children's attentional biases in the context of naming are the product of attentional learning engendered by word learning itself. The present research consists of 11 experiments directed to three objectives: (l) We will track the developmental origins of attentional biases in word learning and their relation to the specific kinds of words that children know in three experiments that examine vocabulary growth and artificial word learning in children from 17 to 24 months; (2) In four studies, we will teach children from 18 to 24 months of age new lexical categories in an effort to teach them generalizable attentional biases than then promote the learning of other words; (3) In four additional experiments involving 18 and 24 month olds, we will contrast the possible special tie between attentional learning and naming by comparing attention in the task of naming to that in other categorization tasks. In the core method, children are presented with novel 3-dimensional objects made of wood, wire, clay, and sand. The object is named. The children then decide whether other objects can be called by the same name. The method duplicates the context of natural word learning.